Understanding Regional Characteristics Through EC Data Analysis

Kohei Yamaguchi, Kazuyuki Shoji, Naoki Tamura, Namiki Sakakura, Yuki Matsukura, Yu Hiranaga, Kohei Shimomura, Kenta Urano, Takuro Yonezawa, Nobuo Kawaguchi

Abstract: The COVID-19 pandemic, which began in 2020, has changed people’s lives, and people are shopping more online. While the analysis of online shopping is becoming increasingly important, regional differences in consumption trends exist. This study proposes a data-driven regional modeling method based on EC purchase data to examine the regional characteristics of online shopping purchase trends. Using the proposed method, we quantified and visualized the degree of similarity of consumption trends among regions and the degree of dispersion of consumption trends within regions using approximately 300,000 lines of online shopping history for 3 years in Japan. As a result, we found that there was some disruption in the early stages of e-commerce for food products by region, while there was little difference in consumption trends among regions for daily necessities. In addition, for consumer durables and clothing, regional differences in consumption trends were confirmed in terms of the number of cars owned per capita, urbanization status, and other regional characteristics.